Performance of a digital subscriber line (DSL) in terms of capacity depends on a number of factors such as attenuation and a noise environment. Since DSL is adaptive to its environment, equipment is necessary to test how well a DSL line will perform in a certain environment.
Performance of a DSL transmission system is impacted by crosstalk interference from one twisted line pair to another twisted line pair with the same binder and, to a lesser extent, twisted line pairs in neighboring binders.
An example of near end crosstalk (NEXT) and far end crosstalk (FEXT) is illustrated in FIG. 1. As shown in FIG. 1, a DSL transmission system 10 may include DSL twisted line pairs L1-Ln in a binder 15. The DSL twisted line pairs L1-Ln may extend from a central office 20 to a customer premises equipment (CPE) side 30. The DSL twisted line pairs L1-Ln may be coupled to the CPE side 30 via a DSL transceiver (not shown). The binder 15 may hold up to 25 or 50 twisted line pairs.
Any one of the lines L1-Ln may be considered a victim line and the remaining lines L1-Ln may be considered disturber lines. For the sake of clarity and brevity, L1 will be described as the victim line.
The victim line is the line in which performance is analyzed. As shown, NEXT is the coupling that occurs between a transmitted signal at the near end side of a disturber line, for example, the twisted line pairs L2-Ln, and a signal at a receiver (not shown) at the near end side of the victim line L1. Furthermore, NEXT may occur at the central office 20 side or at the side of the CPE 30, as illustrated in FIG. 1.
Contrary to NEXT, FEXT occurs when signals from the victim line L1 and signals from the disturber lines L2-Ln become coupled as the signals are sent from the central office 20 to the CPE side 30. Ignoring spectrum allocations and focusing on signal strength, FEXT is typically less harmful because attenuation of the disturber is following the same path as the attenuation of the victim line. NEXT is typically more harmful because the coupling interference has a shorter path to travel whereas the signal from the victim line L1 has a long way to travel.
However, popular deployed DSL systems like Asymmetric DSL (ADSL) and Very High Bitrate DSL (VDSL) have been designed in such a way that upstream and downstream frequency bands do not overlap. Consequently, NEXT is “out of band” and, thus, not harmful. On the other hand, FEXT will impact performance.
Until now, multipair DSL systems were only custom engineered prior to deployment. However, multipair systems are now being deployed in residential markets and, thus, are becoming mainstream. Thus, testing and simulating realistic crosstalk environments has become a more important aspect of the multipair DSL systems.
To simulate crosstalk, some have used a noise generator that injects noise onto the line. This is a common practice for testing DSL systems. The noise that is injected onto the line is typically generated with an arbitrary waveform generator (ARB). The waveform of the noise can be programmed. The programmed waveform will generally follow established noise and crosstalk models as documented in standards such as ATIS T1.417.
However, these implementations are limited to noise models that are put into the noise generator. Furthermore, the noise models are typically applied in a stationary fashion, thereby not allowing for testing dynamic behavior and robustness of DSL technology under non-stationary noise conditions, for example, during startup.
Another way of simulating crosstalk includes using a real cable plant assembled in a series of cable segments connected via instance relay matrices. A DSL system may be connected to such an environment to experience any-to-any pair crosstalk. However, there can be a significant variation in coupling path from one pair to another. For example, in a 25 pair cable, if crosstalk from pair 1 to 2, pair 2 to 3, pair 1 to 3 and so on was measured to form the whole 25-by-25 matrix, there would be 625 crosstalk curves, each differing from the other. Crosstalk within the same cable can also vary from segment to segment. Using such a system as a test environment does not allow for reproducible performance results.